BEFORE WE START

terminal

pip install anthropic requests beautifulsoup4

terminal

# macOS / Linux export ANTHROPIC_API_KEY='your-key-here'# Windows (Command Prompt) set ANTHROPIC_API_KEY=your-key-here

01

Project Structure

~2 minutes

terminal

mkdir ai-research-assistant cd ai-research-assistant touch research_assistant.py

02

Build the Web Content Reader

~5 minutes

🔍

research_assistant.py — Part 1: Web Reader

import os import requests from bs4 import BeautifulSoup import anthropic# Initialise the Claude client client = anthropic.Anthropic( api_key=os.environ.get("ANTHROPIC_API_KEY") )def fetch_webpage_content(url: str, max_chars: int = 8000) -> str: """Fetch a URL and return clean paragraph text only.""" try: headers = {'User-Agent': 'Mozilla/5.0 (ResearchBot/1.0)'} response = requests.get(url, headers=headers, timeout=10) response.raise_for_status()soup = BeautifulSoup(response.text, 'html.parser')# Remove noise: scripts, styles, nav, ads for tag in soup(['script', 'style', 'nav', 'footer', 'header', 'aside']): tag.decompose()# Extract paragraph text paragraphs = soup.find_all('p') text = ' '.join( p.get_text(strip=True) for p in paragraphs if p.get_text(strip=True) )# Trim to token-friendly length return text[:max_chars] + ("... [truncated]" if len(text) > max_chars else "")except Exception as e: return f"Error fetching {url}: {str(e)}"

03

Build the AI Summariser

~5 minutes

research_assistant.py — Part 2: AI Summariser

def summarise_with_claude( content: str, research_question: str, source_url: str ) -> str: """Send content to Claude and get a research-focused summary."""system_prompt = """You are a precise research analyst. Extract and summarise information directly relevant to the user's question. Structure your response as:FINDINGS — 3-5 bullet points of the most relevant facts INSIGHT — 2-3 sentences of your analytical interpretation LIMITATIONS — what this source does NOT cover SOURCE QUALITY — rate High / Medium / Low with one-line reason """user_message = f"""Research question: {research_question}Source URL: {source_url}Content to analyse: {content}Please summarise the above content relative to my research question."""response = client.messages.create( model="claude-sonnet-4-20250514", max_tokens=1024, system=system_prompt, messages=[{"role": "user", "content": user_message}] ) return response.content[0].text

04

Add Conversation Memory

~5 minutes

research_assistant.py — Part 3: Memory

# Conversation history — persists for the whole session conversation_history = []def chat_with_memory(user_message: str) -> str: """Send a message to Claude with full conversation history."""# Append the new user turn conversation_history.append({ "role": "user", "content": user_message })response = client.messages.create( model="claude-sonnet-4-20250514", max_tokens=1024, system="""You are a personal AI research assistant. You maintain full context of this research session and help synthesise findings from multiple sources.""", messages=conversation_history # full history every time )assistant_reply = response.content[0].text# Append Claude's reply to maintain the thread conversation_history.append({ "role": "assistant", "content": assistant_reply })return assistant_reply

05

Multi-Source Synthesis + Save Report

~8 minutes

research_assistant.py — Part 4: Synthesis + Save

def synthesise_sources(summaries: list, research_question: str) -> str: """Combine multiple source summaries into one final report.""" combined = "\n\n---\n\n".join( f"SOURCE {i+1}:\n{s}" for i, s in enumerate(summaries) ) synthesis_prompt = f"""Research question: {research_question}I have gathered summaries from {len(summaries)} sources. Synthesise them into a final research report with:## EXECUTIVE SUMMARY (3-4 sentences) ## MAIN FINDINGS (consolidated bullet points) ## AREAS OF AGREEMENT (what all/most sources agree on) ## CONFLICTING VIEWS (where sources disagree) ## RESEARCH GAPS (what no source adequately covered) ## NEXT STEPS (2-3 actionable follow-up questions)Source summaries: {combined}"""return chat_with_memory(synthesis_prompt)def save_report(report: str, filename: str = "research_report.txt"): """Save the final report to a file.""" os.makedirs("reports", exist_ok=True) path = f"reports/{filename}" with open(path, "w", encoding="utf-8") as f: f.write(report) print(f"\nReport saved to: {path}")

research_assistant.py — COMPLETE FILE FULL CODE

#!/usr/bin/env python3 """ Personal AI Research Assistant Built with Claude API (claude-sonnet-4-20250514) AiBytec.com — Agentic AI Tutorial """import os import requests from bs4 import BeautifulSoup import anthropic from datetime import datetime# Client & memory client = anthropic.Anthropic(api_key=os.environ.get("ANTHROPIC_API_KEY")) conversation_history = []def fetch_webpage_content(url: str, max_chars: int = 8000) -> str: try: r = requests.get(url, headers={'User-Agent': 'ResearchBot/1.0'}, timeout=10) soup = BeautifulSoup(r.text, 'html.parser') for tag in soup(['script', 'style', 'nav', 'footer', 'header']): tag.decompose() text = ' '.join(p.get_text(strip=True) for p in soup.find_all('p')) return text[:max_chars] except Exception as e: return f"Error: {str(e)}"def chat_with_memory(user_message: str) -> str: conversation_history.append({"role": "user", "content": user_message}) response = client.messages.create( model="claude-sonnet-4-20250514", max_tokens=1500, system="You are a personal AI research assistant. You maintain full context of this research session and help synthesise findings from multiple sources into clear, actionable reports.", messages=conversation_history ) reply = response.content[0].text conversation_history.append({"role": "assistant", "content": reply}) return replydef summarise_source(url: str, research_question: str) -> str: print(f" Fetching: {url}...") content = fetch_webpage_content(url) prompt = f"""Analyse this web content for my research question: "{research_question}"Source: {url} Content: {content}Provide: FINDINGS — 3-5 bullets relevant to my question INSIGHT — 2-3 sentences of analysis LIMITATIONS — what this source misses""" return chat_with_memory(prompt)def synthesise_and_save(summaries: list, research_question: str): combined = "\n\n---\n\n".join( f"SOURCE {i+1}:\n{s}" for i, s in enumerate(summaries) ) final = chat_with_memory( f"""Synthesise these {len(summaries)} source summaries into a final report for: "{research_question}"Include: Executive Summary, Main Findings, Areas of Agreement, Conflicting Views, Research Gaps, Recommended Next Steps{combined}""" ) os.makedirs("reports", exist_ok=True) filename = f"reports/report_{datetime.now().strftime('%Y%m%d_%H%M')}.txt" with open(filename, "w", encoding="utf-8") as f: f.write(f"RESEARCH QUESTION: {research_question}\n\n") f.write(final) print(f"\nReport saved to: {filename}") return final# MAIN if __name__ == "__main__": print("Personal AI Research Assistant — AiBytec.com") print("=" * 50)research_question = input("\nEnter your research question: ") print("\nEnter source URLs (one per line, blank line when done):")urls = [] while True: url = input(f" URL {len(urls)+1}: ").strip() if not url: break urls.append(url)print(f"\nProcessing {len(urls)} sources...\n") summaries = [summarise_source(url, research_question) for url in urls]print("\nGenerating final synthesis report...\n") report = synthesise_and_save(summaries, research_question) print("\n" + "=" * 50) print(report)# Follow-up questions with full memory print("\nAsk follow-up questions (type 'exit' to quit):") while True: q = input("\nYou: ") if q.lower() == 'exit': break print(f"\nAssistant: {chat_with_memory(q)}")

$ run it

python research_assistant.pyPersonal AI Research Assistant — AiBytec.com ==================================================Enter your research question: What are the best AI agent use cases in 2026?Enter source URLs (one per line, blank line when done): URL 1: https://example.com/article-1 URL 2: https://example.com/article-2 URL 3:Processing 2 sources... Fetching: https://example.com/article-1... Fetching: https://example.com/article-2...Generating final synthesis report... Report saved to: reports/report_20260323_1430.txt